From time to time, widespread species decline in abundance so much that they appear to be at risk of extinction. Assessing the causes of such declines is a first step towards developing management actions to reverse declines and recover at-risk species. The monarch butterfly is a widespread species, known for its spectacular migration over multiple generations from breeding grounds in northern North America to overwintering grounds in Mexico and California. During the past 37 years, the monarch population that breeds west of the Rocky Mountains and overwinters in coastal California has declined by about 97%, from ~10 million butterflies in the 1980's to ~300,000 in the 2010's. We analyze potential causes of these declines using partial least squares regression (PLSR), a statistical technique for evaluating how multiple correlated variables (in this case, changes in pesticide use, land use, and climate) relate to an outcome of interest (in this case, estimated abundance in overwintering groves and annual population growth rates of monarch butterflies from 1981 through 2017).
Results/Conclusions
In the analysis of monarch abundance, the first PLSR axis explained 64% of the variance in abundance. This multivariate predictor was strongly negatively correlated with increasing land-use intensity, including annual levels of neonicotinoid and glyphosate use, as well as with development of coastal lands near overwintering grounds in California. In the analysis of annual monarch population growth rates, the first PLSR axis explained 30% of year-to-year variation. This multivariate predictor was strongly positively correlated with warm temperatures and high precipitation at the breeding grounds during breeding season, and with cool temperatures and low drought across the western breeding range. For both abundance and population growth rates, the fact that multiple possible causal factors were equally correlated with the multivariate predictor variable means that we cannot separate their effects using monitoring data alone.
In conservation biology, it is often tempting to immediately point to one particular cause of a declining population, and focus on that alone. For western monarchs, we see a broad trend that increasing land use intensity is associated with population declines, and that climate variables may contribute to annual fluctuations in growth rate around these declines. However, likely environmental drivers covary too much in nature to quantify their individual effects, or identify which are causal and which merely correlative. In order to set specific management guidelines, we will need a mechanistic understanding of demography and movement in relation to environmental drivers.